For anyone who wants to do their own analysis, there's an eruptive history of Mayon here.

I pulled 41 more-or-less usable start dates off the list, and typed them into an astronomy simulator, my venerable copy of Dance of the Planets. I was pretty generous with what I called a full moon, allowing for lopsided gibbous phases, maybe three days either side of full. Interestingly, six out of the last seven eruptions on that list came up with this sort of "full-ish" moon, which may account for popular recollection and perception. The count goes down in the long term, though, hitting 13 out of 41 (32%).
Now, the chance of spotting a moon in the phase range I used is only about 20%, suggesting I should have found only eight "full-ish" moons.
However, the sample is too small to say anything useful: the 95% confidence interval on 41 trials with a 20% probable outcome includes the possibility of 13 hits, if I'm doing my sums right.

Grant Hutchison

Disclaimer:: I copied and typed and eyeballed. There may well be some errors in my figures. I hope someone feels the urge to attack these data more scientifically.